streptomyces r61 dd-peptidase

Biochem. J. (1993) 291, 537-544 (Printed in Great Britain)

The mechanism of action of DD-peptidases: the role of tyrosine-159 in theStreptomyces R61 DD-peptidaseJean-Marc WILKIN, Marc JAMIN, Christian DAMBLON, Guo-Hua ZHAO, Bernard JORIS, Colette DUEZ and Jean-Marie FRERE*Centre d'lng6ni6rie des prot6ines and Laboratoire d'Enzymologie, Universite de Liege, Institut de Chimie B6, B4000 Sart-Tilman, Liege, Belgium

Tyrosine- 159 of the Streptomyces R61 penicillin-sensitive DD-peptidase was replaced by serine or phenylalanine. The secondmutation yielded a very poorly active protein whose rate ofpenicillin binding was also drastically decreased, except for thereactions with nitrocefin and methicillin. The consequences ofthe first mutation were more surprising, since a large proportion

INTRODUCTION

The membrane-bound penicillin-binding proteins (PBPs) controlthe biosynthesis of the bacterial cell wall peptidoglycan (Frereet al., 1992). In vitro, some of these proteins exhibit DD-carboxy-peptidase and/or DD-transpeptidase activities. Together with theactive-site-serine ,3-lactamases, they form a superfamily ofpenicillin-recognizing enzymes, and the differences between thetwo types of enzyme can be summarized in a simple sentence:,3-lactams inactivate PBPs, while 8-lactamases inactivate ,3-lactams. Although this statement accounts for most enzyme-fl-lactam interactions, it is an oversimplification since, for example,some 3-lactams are known to also inactivate /3-lactamases.PBPs and 3-lactamases appear to share some important

properties. (1) As stated above, they are active-site-serine en-

zymes, with the exception of some Zn2+-,8-lactamases which willnot be discussed here (Frere et al., 1976a; Waley, 1992). (2)Kinetically, their interaction with ,3-lactams can most often berepresented by the simple, 3-step model shown below (Frereet al., 1975; Christensen et al., 1990).

k+1 k+2 k+3(H20)E + C --E- C E-C* E + P(s) (model 1)

k-I

where E, C, E * C, E-C* and P(s) are respectively the enzyme, the,3-lactam, the non-covalent Henri-Michaelis complex, the acyl-enzyme and the reaction product(s), which are devoid of bio-logical activity. (3) Some esters and thiolesters behave as

substrates for both DD-peptidases and ,-lactamases (Adam et al.,1990, 1991), although the latter enzymes fail to recognize thenormal peptide substrates of the former. (4) X-ray diffractiondata have indicated clear similarities in the three-dimensionalstructures of several 8-lactamases and of the Streptomyces R61extracellular DD-peptidase (Kelly et al., 1986; Samraoui et al.,1986), and sequence alignments have allowed the identificationof several conserved elements (Joris et al., 1988), which are

probably directly involved in the reaction mechanism (such as

the active-site serine) or bordering the catalytic cavity.Besides the specific interactions of the DD-peptidases with the

D-alanyl-D-alanine-terminated peptides, the most striking

of the thiolesterase activity was retained, together with thepenicillin-binding capacity. Conversely, the peptidase propertieswas severely affected. In both cases, a drastic decrease in thetransferase activity was observed. The results are compared withthose obtained by mutation of the corresponding residue in theclass A ,-lactamase of Streptomyces albus G.

difference between the two types of enzyme is the value of k.,which is generally quite high for f8-lactamases (up to 4000 s-')and extremely low for DD-peptidases (10-3-10-6 s-1). Thisdifference is responsible for the 'macroscopic' consequencesdescribed above.One of the conserved elements is a Ser-Asp-Asn sequence

situated on a loop near the active-site serine of class A ,3-lactamases (Joris et al., 1991). The corresponding triads are Tyr-Ala-Asn and Tyr-Ser-Asn in class C,/-lactamases and in the R61DD-peptidase respectively (Ghuysen, 1991). The roles of thesethree residues in the catalytic properties of class A ,3-lactamaseshave been extensively studied by site-directed mutagenesis (Jacobet al., 1990a,b), and the ionized group of the tyrosine residue ofclass C ,3-lactamases has been hypothesized to be the generalbase which activates the active serine by abstracting its proton(Oefner et al., 1990).

In this paper, we have assessed the importance of the cor-responding tyrosine residue (Tyr-159) in the Streptomyces R61DD-peptidase by site-directed mutagenesis. This enzyme is by farthe most studied penicillin-sensitive DD-peptidase and the onlyone for which three-dimensional data are available.

KINETIC MODELThe interaction between ,l-lactams and DD-peptidase is wellrepresented by model 1 (Frere et al., 1975), with the first step inrapid equilibrium (K = k.1/k,,). When [C] << K, the rate ofacylation is characterized by a second-order rate constant, k+2/K,which is most often the parameter responsible for the sensitivityof the enzyme to the inactivator.Model 1 also depicts the carboxypeptidation pathway where

a first product, P1, is released in the k+2 step. With the bestpeptide substrate, NaNN-bisacetyl-L-IYSVI-D-alanyl-D-alanine(Ac2KAA), k+2 is < k+3 and thus kcat. = k2 and Km= K'=(kW1 +k+2)k-1 (Varetto et al., 1987); P1 is D-alanine. Conversely,with the thiolester benzoylglycylmercaptoacetate (C6H5-CO-NH-CH2-CO-S-CH2-COO-), k+3 is < k+2; thus kcat. = k+3 andKm = k+3K'/k+2; P1 is mercaptoacetate (Jamin et al., 1991).

In the presence of suitable aminated acceptors, the enzymealso catalyses the transfer of the Ac2-L-lysyl-D-alanyl or benzoyl-

Abbreviations used: Ac2KAA, /\N6-bisacetyl-L-IYSYI-D-alanyl-D-alanine; AC2KALa, /VNf-bisacetyl-L-lysyl-D-alanyl-D-lactate; Ac2KATI, Nat'P-bisacetyl-L-lysyl-D-alanyl-D-thiolactate; PBP, penicillin-binding protein; BPA, benzylpenicilloic acid.

* To whom correspondence should be addressed.

537

538 J.-M. Wilkin and others

glycyl moieties of the donor substrates on to the acceptor aminogroup (transpeptidation). With the thiolester, alcohols can alsoserve as acceptors in transesterification reactions. When k+3 is< k+2, the presence of an acceptor increases the rate of donorsubstrate disappearance by opening a new branch for theutilization of the acyl-enzyme, E-C*. However, a simple partitionmodel is not sufficient to account for all of the experimentalobservations (Frere et al., 1973; Jamin et al., 1991).

MATERIALS AND METHODS

ChemicalsEnzymes for genetic engineering were purchased from Biolabs(Beverly, CA, U.S.A.) and Boehringer (Mannheim, Germany).[35S]dATP (1000 Ci/mmol) was from Amersham; benzyl-penicillin was from Rhone-Poulenc (Paris, France), [14C]benzyl-penicillin (53 Ci/mol) from Amersham, carbenicillin fromBeecham Research Laboratories (Brentford, Middlesex, U.K.),ampicillin from Bristol Benelux (Brussels, Belgium), cephalo-sporin C from Elli Lilly and Co. (Indianapolis, IN, U.S.A.) andcefuroxime from Glaxo Group Research (Greenford, Middlesex,U.K.). All of the unlabelled ,-lactam compounds were kindlygiven by the respective companies. Nitrocefin was purchasedfrom Oxoid (Basingstoke, Hampshire, U.K.).The tripeptides Ac2KAA and NaNe-bisacetyl-L-IYSYI-D-alanyl-

D-thiolactate (Ac2KATl) were from UCB Bioproducts (Braine-l'Alleud, Belgium). N"N"-Bisacetyl-L-IYSYI-D-alanyl-D-lactate(Ac2KALa) was a gift from Dr. R. F. Pratt (Middletown, CO,U.S.A.). The other substrates were synthesized in our laboratory.The structures of the various substrates are given in Table 1.

Rabbit anti-(R61 DD-peptidase) antiserum was from GammaS.A. (Angleur, Belgium).

OligonucleotidesOligonucleotides were a gift of Dr. J. Brannigan, University ofSussex, Brighton, U.K. The crude oligonucleotides were purifiedwith the help of an Oligonucleotide Purification Cartridge(Applied Biosystems).The oligonucleotides used to introduce the mutations were the

following: Tyr-1 59 -+ Phe, GGCCGCGCCTATTCATTCTCC-AACACGAAC; Tyr-159 -. Ser, GGCCGCGCCTATTCAT-CCTCCAACACGAAC; the modified bases are underlined (thewild-type codon was TAC).

Table 1 Structures of the substratesThe general formula is given by:

R'-NH-CH-CO-X-CH-COO-

R2 R3

Substrate R' R2 R3 X

Ac2KAAAc2KALaAC2KATISleS2aS2cS2dS2eS2VaI

AC2L-LysAc2L-LysAc2L-LysC6H5-COC6H5-COC6H5-COC6H5-COC6H5-COC6H5-CO

CH3 (D)CH3 (D)CH3 (D)HHHCH3 (D)CH3 (D)(CH3)2-CH3 (DL)

CH3 (D)CH3 (D)CH3 (D)C6H5-CH2HCH3 (D)HCH3 (DL)H

Strains, plasmids and molecular biology techniques

Escherichia coli strain TGl was used as the host for phage Ml3and Streptomyces lividans strain TK24 (Hopwood et al., 1985)was used for enzyme expression and production.The multicopy Streptomyces plasmid pDMLI15 was the

expression vector. It carries the gene of the exocellular DD-peptidase of Streptomyces R61 (dacR61) and presents uniqueSphI and PstI sites (Duez et al., 1992).The recombinant procedures were those described by Maniatis

et al. (1982) and Hopwood et al. (1985), except that plasmid andM13 replicative form extractions were performed using theQIAGEN protocols and Tip-100 or Tip-20 columns (QIAGEN,Studio City, CA, U.S.A.). The Streptomyces plasmid extractionswere performed as described by Hopwood et al. (1985), but thecentrifugation step was replaced by a QIAGEN protocol. Elutionof DNA fragments from agarose gels was carried out with thehelp of a Geneclean kit (Bio 101, La Jolla, CA, U.S.A.).

In order to introduce oligonucleotide-directed changes, thePstI/SphI fragment was subcloned in phage M13mpl9 to providesingle-stranded template DNA. The replacement of the tyrosinecodon (TAC) by that for phenylalanine (TTC) or serine (TCC)was performed using the procedure of Taylor et al. (1985), withthe help of an oligonucleotide-directed in vitro mutagenesis kit(Amersham International). The clones were screened bysequencing the complete fragment subcloned in M13. Thedideoxy method was used with the T7 sequencing kit (PharmaciaLKB Biotechnology AB, Uppsala, Sweden). The mutatedfragments were re-inserted into pDML1 15 to reconstitute thewhole gene and the plasmids thus obtained were used totransform protoplasts of Streptomyces lividans TK24. Thetransformants were screened by immunodetection (Bio-RadImmunoblot Assay kit). The absence of any other mutation wasconfirmed by resequencing the mutant DD-peptidase gene usingthe oligonucleotides priming at regular intervals of 150 bp.

Production of the mutant DD-peptidasesFor the expression of the mutant proteins, transformantStreptomyces colonies selected on thiostrepton-containing R2YEagar (Hopwood et al., 1985) were used to inoculate a 250 ml pre-culture in modified YEME medium (Erpicum et al., 1990).Thiostrepton (25 mg/l) was added to exert a selection pressureduring the exponential growth phase, since the plasmid carriedthe resistance to this antibiotic. Baffled 1 litre Erlenmeyer flaskswere used and incubations were at 28 °C with orbital agitation at250 rev./min (New Brunswick Scientific incubator). After 3 days,a 4% inoculum was used to initiate the main culture, which wascontinued for 5 days under the same conditions.

Purification of the mutant Do-peptidasesThis was performed as described by Hadonou et al. (1992). Insummary, complete purification required three or four steps afterthe elimination of the mycelium by centrifugation: (1) adsorptionat pH 4.0 on Amberlite CG-50 and desorption at pH 8.0; (2)chromatography on Q-Sepharose Fast Flow; (3) filtration onSephadex G-75; (4) chromatography on a prepacked Hiload26/10 Q-Sepharose Fast Flow column. The last step has not beendescribed before. The buffer was 10 mM Tris/HCl containing50 ,uM EDTA, pH 7.0, and a salt gradient up to 0.2 M NaCl wasapplied over 1.1 litres.The enzyme was detected in the various fractions by

SDS/PAGE using the wild-type Streptomyces R61 enzyme as amolecular mass standard.

Mechanism of action of DD-peptidases 539

Kinetic parameters

SubstratesSeveral substrates (see Table 1) could be directly and continuouslymonitored at 250 nm with the following absorbance variations:Sle (Ac 540 M-l cm-') S2a, S2c, S2d, S2e, S2-Val (Ac -2200M-l cm-') and Ac2KATI (Ac -2000 M-' cm-'). Hydrolysis ofthiolester substrates could also be monitored at 324 nm in thepresence of 2 mM 4,4'-dithiodipyridine (Sigma Chemical Co., St.Louis, MO, U.S.A.), which reacts with the thiol groups releasedduring the reaction (Ac 20000 M-l cm-') (Kelly et al., 1992).

Hydrolysis of the tripeptides Ac2KAA and Ac2KALa wasmonitored by estimating the release of D-alanine (D-amino acidoxidase procedure; Frere et al., 1975) or of hydrolysed productAc2KA (see h.p.l.c. procedure) respectively.The steady-state parameters kcat, Km and kcat /Km were de-

termined either from complete time courses (method A; DeMeester et al., 1987) or from initial rates using the Haneslinearization of the Henri-Michaelis equation (method B). Thislatter procedure was always utilized for the tripeptide and thedepsipeptide.

,/-LactamsOn the basis of model 1, fl-lactams could be used as eitherinactivators or substrates. In the first case, the characteristicconstants of the acylation (k+2, K' or k+2/K') and deacylation(+k3) reactions were determined separately using one of thefollowing methods. (1) The apparent first-order inactivation rateconstant ki was determined either by recording the decrease inprotein fluorescence at 320 nm (method la) (excitation at 280 nm;Frere et al., 1975) or by the reporter substrate method (DeMeester et al., 1987) with S2a or Ac2KATl as reporter substrates(method lb). When possible, the individual values of k+2 and K'were determined by fitting the measured ki values to the equation:

ki = k+2[C]/(K' + IC])with the help of the ENZFITTER program (Leatherbarrow,1987). (2) k+2/K' values could also be obtained by the competitionmethod, with nitrocefin as reporter inactivator (Frere et al.,1992). Complete acylation of the enzyme by nitrocefin resulted inan increase in the solution absorbance at 482 nm (AA.); whenthe enzyme was added to a mixture of nitrocefin (N) and asecond fl-lactam (C), this increase (AA1) was smaller, since aproportion of the enzyme was acylated by the other antibiotic;it can be shown that:

AAo-AAi (k+2/K')c [C]

AAi (k+2/K')N [N]where (k+2/K')c is the only unknown. This procedure could onlybe used if acylation reactions were completed over a time periodmuch shorter than the half-lives of both acyl-enzymes. (3) Therate of deacylation (k+3) was determined by monitoring the re-activation of the enzyme after elimination of the excess free ,-lactam by addition of a small amount of Bacillus licheniformis f/-lactamase. To determine the recovered activity, the tripeptideAc2KAA was used after inactivation by compounds exhibiting alow k+3 value (10-5-10-6 s-'; most cephalosporins), and thethiolester substrate S2a was also used after inactivation bycompounds exhibiting a higher k+3 value (10-3-10-4 s-'; nitrocefinand penicillins). Identification of the degradation products of the['4C]benzylpenicilloyl-enzyme adducts was performed by t.l.c.(Frere et al., 1976b).When 8-lactams were used as substrates (method 4),

the time courses exhibited a burst, followed by steady-state

linear hydrolysis. Analysis of the burst yielded a k valueI = k+3+ [k+2[C]/(K' + [C])} and the steady-state rate was equal tok+3[Eo], assuming that k+3 was much smaller than k,{k+2[C]/K' + [C])}, a condition that was verified in all cases. Whenki varied linearly with [C], only k+2/K' could be computed. Thisprocedure thus allowed a rapid determination of the acylationand deacylation parameters. The time courses were recorded at482 nm (nitrocefin), 260 nm (other cephalosporins) or 230 nm(penicillins).

All kinetic experiments were performed at 37 °C (unlessotherwise stated) with a Uvikon 860 or Hewlett Packard 8452Aspectrophotometer coupled to microcomputers via RS232interfaces, or with a Perkin-Elmer MPF44 spectrofluorimeter.

Reactions between substrate S2a and mutant DD-peptidasesunder pre-steady-state conditions were monitored with a BiologicSFM3 stopped-flow apparatus. As shown before (Jamin et al.,1991) the fluorescence of the acyl-enzyme intermediate was lowerthan that of the free enzyme, and quenching of the fluorescenceemission at 20 °C was recorded through a 310-490 nm band filterwith excitation at 290 nm.

Denaturation experimentsDenaturation of the Streptomyces R61 DD-peptidase induced ashift in the fluorescence emission maximum from 320 nm to340 nm (excitation 280 nm) (Nieto et al., 1973). The quenchingof fluorescence at 320 rm obeyed first-order kinetics, and thedenaturation first-order rate constant (kd) was computed withthe help of the following equation:

ln(Ft-F0o) = ln(F0-F,,)-kdtwhere Fo, Ft, F, were the fluorescences at times zero, t andinfinity respectively.Two sets of denaturation conditions were used to characterize

the stability of the mutant enzyme: 10 mM sodium phosphatebuffer, pH 7.0, at 60 °C, or 6 M urea in the same buffer at 56 'C.

H.p.i.c. proceduresChromatography experiments were performed on a Nucleosil 5-C,8 (Machery-Nagel) column (ET250/8/4) coupled to a Merckh.p.l.c. apparatus. Separation of Ac2KALa and its hydrolysisproduct, Ac2KA, was performed under isocratic conditions using10 mM sodium acetate/acetic acid buffer, pH 3.0, containing3 % acetonitrile. The retention times were 8.6 and 5 min for thesubstrate and the product respectively. For transpeptidationreactions with S2a, the conditions were as described by Jamin etal. (1991). Quantification was achieved by integration of the peakareas using pure Ac2KALa, Ac2KA, hippuric acid or S2a asstandards.

RESULTSMutagenesis, production and purificationThe mutagenesis procedures yielded two modified plasmids,pDML35 (Tyr-159 - Phe) and pDML36 (Tyr-159 -+ Ser). Theplasmids were purified and the modified genes were sequenced;no additional mutations were detected.Enzyme production was performed in the optimized YEME

medium (Erpicum et al., 1990). Maximum production wasobtained after 5 days of growth, as observed for the wild-typeprotein. This yielded 120 and 45 mg of modified enzyme per litrefor the Tyr-159 -+ Phe and Tyr-159 -. Ser proteins respectively.The purifications are summarized in Table 2. Step 4 allowed

the purification of two enzyme populations of similar Mr(37 000-38 000), but which were eluted at different ionic strengths.


Table 2 Purffication of the Tyr-159 mutant proteins

Culture volumes were 2.25 litres for both mutant enzymes. S.D.s were < 10% of means.

Enzymes ... Tyr-1 59 -* Phe Tyr-1 59 -. Ser

Total Total Purity Total Total PurityStep protein (mg) enzyme (mg) (%) protein (mg) enzyme (mg) (%)

234

6400460200100

300200150100

4.7 260043 27575 100

>95 -

100100100

3.836

> 95

Table 3 k,L, Km and k.LIK, values for the wild-type and Tyr-159-substituted mutant enzymes

Values for the wild-type enzyme are from *Frere and Joris (1985), tVarebto et al. (1987) or lAdam et al. (1990). M is the method used to calculate the steady-state parameters: A, completetime course method (De Meester et al., 1987); B, linearization of the Henri-Michaelis equation (see the Materials and methods section). S.D.s did not exceed 15% of the means. ND, not determined.

Tyr-1 59 -+ Phe Tyr-1 59 -+ Ser Wild type

kcat Km k/aIK kw Km k /K kcat Km k/KmSubstrates (s-1) (#M) (M-1 .s') M (s.1) (jM) (M-1 s-1) M (s-1) (M) (M-1 .s1) M

0.5 B 0.02 13 000 1.3 B1.4 B 8 45000 180 B

30 4000 9000 B

< 1 B 0.02 1400 14 A00 B 0.3 5 60000 A

0.22 2 114000 A)0 B 9 50 200000 A

11 200 52000 A0.46 100 4500 A

55* 14000 400032t 40000 80072 8000 900051 900 550051 50 10000051 50 100000

70 100 70000070 560 800004 500 8000

140(b)

I- I T

af E

6.5 8.5 0--54.56.5 8.5 10.5 4.5

8

a4i

0I; 6.5 8.5 10.5 4

pH.5 6.5 8.5 10

pH

U)

ME 150-

x100

.5 4,100

en

i 50X

x

04).5 4.

(c)

,5

Figure 1 pH-dependence of the kinetic parameters for the hydrolysis of substrate S2a by the wild type (0) and Tyr-159 -- Ser (0) enzymes: (a) k.L values,

(b) Km values, (c) kpLIKm values

The values were obtained by analysing complete time-courses. Reactions were performed at 37 OC in the following buffers: pH 5.0, 10 mM sodium cacodylate/HCI; pH 6.0, 13 mM sodium

cacodylate/HCI; pH 7.0, 10 mM sodium phosphate; pH 8.0, 7 mM sodium phosphate; pH 9.0, 6 mM potassium borate/NaOH. The ionic strength was adjusted to the same value throughout

by the addition of NaCI. The pH profile of the wild-type enzyme is from Varetto (1991). Error bars represent the S.D.s. (n = 6 or more).

Ac2KAAAc2KALaAc2KATLaSleS2aS2cS2dS2eS2Val

0.006 13000ND NDND NDND ND0.008 70ND ND0.05 50ND NDND ND

ND

1(ND100NDND

(a)

BBBAAAAAA

T

T6

.;4w

2

4.5

T

a

U)_0.2

I Iii

I I II T

10

6.5 8.5 10

O L4.5

6.5 8.5 lC

I

I I

1.5

IiII

6.5 8.5

pH10.5

.5


Table 4 Influence of the acceptor concentration on the k,,t,, K, and k,,LlKm values for the wild-type and mutant enzymes

S2a was the donor substrate. M is the method used to calculate the steady-state parameters: A, complete-time course method (De Meester et al., 1987); B, linearization of the Michaelis equation.S.D.s did not exceed 15% of the means. ND, not determined.

Tyr-1 59 -* Phe Tyr-1 59 -* Ser Wild type

kcat Km kaI/K kt Km k /K kcat Km k/aIKAcceptor (s-1) (1tM) (M-1 . s-1) M (s.1) (1tM) (M-1 . s-1) M (s-1) (#M) (M-l . s-1) M

0.008 70 100 B 0.30 5.0 60000

ND ND ND0.011 100 110ND ND ND0.013 83 150

0.29 5.0 58000B 0.26 5.2 50000

0.25 4.4 57000B ND ND ND

0.006 40 150 B 0.25 6.0 42000

A 5.0 50 100000 A

A 17.0 170 100000 AA 29.0 290 100000 AA 61.0 610 100000 A

67.0 670 100000 A

A 9 90 100000 A

Table 5 Transpeptidation/hydrolysis ratios for the wild-type and mutantenzymesTranspeptidation was with S2a as donor substrate (250 1uM) and D-alanine as acceptorsubstrate.

Transpeptidation/hydrolysis ratio[D-Alanine](mM) Tyr-159 -- Phe Tyr-159-+Ser Wild type

5 0.01 + 0.00510 0.16+0.0215 0.13+0.0220 0.21 +0.0325 -

50 0.19+0.02100 0.25+0.04150 -

200 -

2.7+ 0.1

0.07 + 0.020.20 + 0.050.33 + 0.070.40 + 0.08

4.3 + 0.66.0 + 0.8

The first (70% of the total) was eluted with 73 mM NaCl, andthe second (30 %) with 85 mM NaCl. Both variants hydrolysedsubstrate S2a with apparently identical kinetic parameters withinthe limits of experimental error, and were thus considered to bekinetically equivalent. This phenomenon, which was observedpreviously with the wild-type enzyme and other mutants(Bourguignon-Bellefroid et al., 1992), is under currentinvestigation.The purity of the mutant proteins was verified by SDS/PAGE

and evaluated to be greater than 9500.

Stability and physical propertiesAbsorption spectra of the modified proteins between 230 and320 nm were superimposable on that of the wild-type enzyme.The fluorescence emission spectrum of the Tyr- 159 -+ Phe mutant

was also identical to that of the wild-type enzyme but, with theTyr- 159 -+ Ser protein, a small blue shift ofthe maximum (317 nminstead of 319 nm) was observed, indicating a slight modificationof the fluorophore environment. No major change in stabilitywas observed in the mutant proteins. The half-lives were 6.2 + 0.6(wild type), 6.0+0.6 (Tyr-159 -. Phe) and 3.4+0.3 (Tyr-159 -+

Ser) min at 60 'C in buffer, and 8.6 + 0.9 (wild type), 2.3 + 0.2(Tyr-159 Phe) and 2.3 +0.2 (Tyr-159 -+ Ser) min at 56 °C inthe presence of 6 M urea.

Kinetic properties

Carboxypeptidase and esterase activities

Table 3 compares the substrate profiles of the wild-type andmutant enzymes. Stopped-flow experiments, in which the ac-

cumulation of acyl-enzyme was monitored by recording the rateof fluorescence decrease at various substrate concentrations,allowed the calculation of the individual k+2 and K' values at20°C: Tyr-159-+Phe, k+2 =2.6+0.5s-', K'= 17+3mM;Tyr-159-.Ser, k+2 =300+50s-1, K'= 14+2mM; wild type,k+2 =200+50 s-1, K' = 5+0.5 mM (Jamin et al., 1991).

These results showed that, with substrate S2a, deacylationremained rate-limiting (k+3 < k+2) for the mutants, as found forthe wild-type enzyme. The same situation prevailed with sub-strates S2c, S2d and S2e, for which fluorescence quenchingexperiments permitted the visualization of the accumulation ofthe acyl-enzyme at [S] > Km.The pH-dependence of the kinetic parameters was only studied

in the most favourable case, i.e. the Tyr-1 59 -- Ser-S2a inter-

actions. The results are presented in Figure 1.

Transpeptidase activity

The influence of two acceptors, D-alanine and glycyl-glycine, on

the kinetic parameters for the utilization of substrate S2a isshown in Table 4. The most striking observation was that, insharp contrast to the situation with the wild-type enzyme, thepresence of an acceptor failed to increase the kcat and Km valuesfor the mutants, indicating that the transpeptidation pathwaywas grossly impaired. Accordingly, only a small proportion oftranspeptidation product could be detected by h.p.l.c. (Table 5).

Interactions with fl-lactam antibiotics

The results presented in Table 6, show drastic decreases in all ofthe acylation rates for the Tyr- 159 -. Phe mutant compared with

the wild-type enzyme. With the Tyr- 159 -. Ser mutant, the k+2/K'value was not significantly decreased. In fact, they were even

increased in several cases. For carbenicillin as substrate, in-dividual k+2 and K' values were determined. They were

0.50 + 0.08 s-' and 0.18 + 0.03 mM respectively, not very differentfrom those observed with the wild-type enzyme (0.10 s-' and0.11 mM). The mutations did not seem to affect the k+3 values.However, at pH 7.0 the rate-limiting step in the degradation ofthe benzylpenicilloyl-wild-type-enzyme adduct is the hydrolysisof the C5-C6 bond of the antibiotic moiety, rapidly followed by

NoneD-Alanine

5.7 mM15 mM45 mM55 mM

Gly-Gly20 mM


Table 6 k+21K' and k+3 values for several antibiotics with the wild-type and mutant enzymes

Determinations of the k+2/K1 and k+3 values were performed as described in the Materials and methods section. Numbers in parentheses refer to the various procedures. S.D.s did not exceed15%; ND, not determined.

Tyr-1 59 -+ Phe Tyr-1 59 -+ Ser Wild type

Antibiotic k.2/K' (M1 s-') ..3 (S...1) k+2/K' (M-1 s1) k..+3 (S1) k+2/K' (M- s-1) k+3 (S-1)

BenzylpenicillinCarbenicillinAmpicillinMethicillinNitrocefinCephalosporin CCefuroxime

55 (2)< 5 (2)< 5 (2)10 (la)

4500 (4)1 4 (2)< 5 (2)

2.1 x 1 0-4 (3)NDNDND4.6 x 10-4 (3)NDND

10 000 (lb)2600 (la)740 (la)

ND5000 (la)540 ( a)740 (1 a)

3.0 x 1 0-' (3)3.3 x 1 0-4 (3)5.9 x 1 0-4 (3)ND4.3 x 10-4 (3)5.0 x 10-' (3)6.6 x 10-u (3)

14000 (la)800 (la)110 (la)10 (la)

4100 (la)1500 (la)350 (la)

1.4 x1 0-4 (3)1.4 x 10-4 (3)1.4 x 1 0-4 (3)ND3.0 x 10- (3)1.0 x 10-u (3)4.0 x 1 0-6 (3)

the hydrolysis of the phenylacetylglycyl-enzyme thus formed(Frere et al., 1976b). Surprisingly, under the same conditions, thedegradation of the adducts formed with the mutants mainlyyielded benzylpenicilloic acid (BPA). The BPA/phenylacetyl-glycine ratios were 50:1 and 10:1 with the Tyr-159 --+ ihe andTyr- 159 -- Ser mutants respectively.

DISCUSSIONAt present, high-resolution refined structures are only availablefor three class A and one class C fi-lactamases. Unfortunately,the X-ray data on the Streptomyces R61 DD-peptidase are stillrather incomplete. In class A enzymes, the catalytic cavity islimited on one side by the Lys-234-Thr(Ser)-Gly motif, cor-

responding to similar Lys-Thr-Gly and His-298-Thr-Glysequences in the class C f-lactamases and the Streptomyces R61DD-peptidase respectively. The highly conserved Ser-130-Asp-Asn loop forms another boundary of the class A enzyme activesite, and site-directed mutagenesis studies have demonstrated theimportant role of these three residues in the stability and catalyticproperties of these enzymes. The corresponding loop is found inthe class C (Tyr-180-Ala-Asn triad) and R61 (Tyr-159-Ser-Asntriad) enzymes. If the other conserved elements [the active-siteserines and the Lys(His)-Thr-Gly triads] are superimposed, thehydroxy group of Ser- 130 (class A /3-lactamases), Tyr-180 (classC /l-lactamases) and Tyr- 159 (R61 DD-peptidase) are found invery similar positions. In consequence, this group could beexpected to take an active part in the catalytic and penicillin-binding properties of the DD-peptidase, a hypothesis that was

fully confirmed by the present study, although with some rathersurprising observations. By using site-directed mutagenesistechniques, the Tyr- 159 residue was replaced by either Phe or

Ser. The first mutation eliminated the hydroxy group while it wasconserved in the second mutation, probably 0.35-0.40 nm furtheraway from the active-site-serine oxygen.The stabilities of the modified proteins were not severely

decreased (never more than 3.7-fold), whereas the Ser- 130 -+ Ala

and Ser- 130 Gly mutants of Streptomyces albus G ,8-lactamasewere 36- and 16-fold more unstable than the wild-type enzyme(Jacob et al., 1990b). Thus no gross modifications of the proteinstructures were suspected, although the small blue shift in thefluorescence emission maximum of the Tyr- 159 -+ Ser mutant

suggested a slightly more hydrophobic environment for thetryptophane residues responsible for enzyme fluorescence. Thefluorescence emission was quenched upon formation of the acyl-enzyme with substrates and f3-lactam inactivators, a situationsimilar to that found with the wild-type enzyme.

The rates of substrate utilization and penicillin inactivation ofthe Tyr-1 59 -+ Phe mutant were drastically decreased, except forthe reactions with nitrocefin and methicillin. The particularbehaviour of these compounds was reminiscent of the situationwith the Ser- 130 -+ Ala and Ser- 130 -* Gly mutants of the S.albus G enzyme, which hydrolysed these two substrates signi-ficantly more efficiently than most other ,-lactams (Jacob etal., 1990b). In the interaction with the thiolester substrate S2a,which could be studied in more detail, the acylation anddeacylation rate constants were severely impaired, by 100- and1000-fold respectively. The loss of catalytic efficiency was evenmore striking with the peptide substrate, for which the kcat value,probably corresponding to k+2, was only 0.01 % of that of thewild type. Conversely, although the Tyr- 159 -+ Ser mutant alsobecame a very poor peptidase, its esterase activity was somewhatbetter, and the thiolesterase and penicillin-binding propertieswere nearly identical to those of the wild type (and sometimesbetter). With substrate S2a and carbenicillin, for which all of theindividual constants were measured, the values were very close tothose observed with the wild-type enzyme.The behaviour of the three substrates containing the Ac2-L-

Lys-D-Ala acylating moiety was particularly striking. The ratesof acylation of the Tyr- 159 -* Ser mutant were respectively 130-and 7000-fold larger with the lactate and thiolactate derivativesthan with the peptide, thus directly reflecting the 'quality' of theleaving group, in contrast to the properties of the wild-typeenzyme. These results thus suggested a role for the Tyr-159hydroxy group in the protonation of the substrate's leavinggroup similar to that proposed for the hydroxy group of Ser- 130in the mechanism of action of the S. albus G ,.-lactamase(Lamotte-Brasseur et al., 1991). Indeed, the pKa values increasefrom 7.0 with the thiolactate to 13-14 with the lactate and to> 14 with alanine. Although a similar explanation seems to holdfor the other thiolesters, for which acylation of the Tyr-1 59 -+ Sermutant appeared to be unimpaired, it does however consitute anoversimplification, since the acylation of the Tyr- 159 -* Phemutant by thiolesters was severely impaired (about 1000-fold).Moreover, with thiolesters containing the benzoyl group, theTyr-159 side-chain also appeared to be involved in deacylation,since the k+3 value decreased by one and three orders ofmagnitudefor the Ser and Phe mutants respectively. In the absence ofdetailed structural information, it seems somewhat dangerous toassume that the serine hydroxy group can replace that of thetyrosine, even with a decreased efficiency. In fact, the ability ofthe Tyr-1 59 -+ Ser mutant to perform some, but not all, of thefunctions of the wild-type enzyme remains a complete surprise.Similarly, acylation of the Tyr-1 59 -+ Ser mutant by fl-lactams


was unimnpaired and was sometimes even faster than with thewild-type enzyme, contrasting again with the dramatic loss ofpenicillin-binding capacity of the Tyr-1 59 -* Phe mutant. Thusby an unknown mechanism such as a structural adjustment orthe involvement of an additional water molecule, the serinemutant could somehow compensate for the modified position ofthe hydroxy group. However, this compensation was onlyefficient with the thiolester substrates and ,J-lactam compounds,indicating more stringent structural requirements for the hy-drolysis of peptide substrates. It could be hypothesized that thethiolester substrates better mimicked the behaviour of ,8-lactamsthan the peptide substrates, but this was somewhat incontradiction with the results obtained with other mutants(His-298 -+ Lys and His-298 -+ Gln), for which the mutationaffected acylation by the ,-lactams 4-20-fold more strongly thanthat by the thiolester S2a (Hadonou et al., 1992).

These results are also reminiscent of those obtained with theLys-213 -. Arg mutant of the E. coli PBP5, which seemed to loseits peptidase activity completely while retaining its penicillin-binding properties (Malhotra and Nicholas, 1992).The most striking result, however, was the near-complete loss

of transpeptidation activity exhibited by both mutants, asdemonstrated by the failure of good acceptors to increase kcatvalues for the thiolester substrate and by the extremely lowtranspeptidation/hydrolysis ratios recorded, even at high D-alanine concentrations. Similar, although less drastic, decreasesin transpeptidation efficiency were also reported for His-298mutants. By analogy with known class A ,f-lactamase structures,and on the basis of the available X-ray data, an interactionbetween Tyr-1 59 and His-298 could be suggested. These twoside-chains must somehow be involved in the interaction with theacceptor substrate, and might co-operate in the catalysis of thetransfer step. Indeed, modifications of these residues resulted inboth cases in a disproportionate decrease in transferase activitycompared with hydrolysis. For instance, with the Tyr-1 59 -+ Sermutant and substrate S2a, the rate of hydrolysis and aminolysis(by 200mM D-alanine) were decreased 17- and 2500-fold re-spectively.The pH-dependence of the kcat and Km values of the Tyr-

159 -+ Ser mutant was seemingly characterized by the appearanceof a new pK of around 6-6.5. This might correspond to theshifting to higher pH values of a pK which is below 4.5 with thewild-type enzyme, and thus outside the pH range which can beeasily studied since the enzyme becomes unstable at low pH. Theinvariance of the kcat./Km ratio in the same pH range suggesteda specific influence on k+3. Conversely, a pK of about 8.5, whichseemed to influence the acyl,.tion of the wild-type enzyme by S2a,was not found with the mutant, and it was thus very tempting toattribute this pK to the tyrosine side-chain. It is, however, quitedangerous to accept such a simple explanation, since the attri-bution of pK values to specific side-chains in an enzyme activesite appears to remain a difficult problem.

Finally, the nature ofthe benzylpenicillin degradation productsrevealed another rather subtle function of the Tyr-1 59 side-chain. With the wild-type enzyme at pH 7.0, the major product(>90%) was phenylacetylglycine, formed by a rate-limitinghydrolysis of the C,{6 bond of the antibiotic moiety, followedby a rapid decay of the phenylacetylglycyl-enzyme thus obtained.In more basic conditions, benzylpenicilloate was also obtained,probably by a direct attack by OH- ions. With the mutants atneutral pH, direct hydrolysis became predominant, and it couldbe calculated that the efficiency of this reaction increased morethan 20-fold while that of the rupture of the C5-C6 bonddecreased 5-20-fold. Thus the mutations seemed not only toallow a direct access of water molecules or OH- ions to the acyl-

enzyme ester bond, but also to impair the more complexdegradation pathway in which Tyr-1 59 might thus play adetermining role.

In conclusion, although several observations remainsurprising, it is clear that the Tyr- 159 side-chain contributes in animportant way to the carboxypeptidase and transpeptidaseactivities of the Streptomyces R61 DD-peptidase. Its influence onthe thiolesterase and penicillin-binding properties is more com-plicated to assess in the light of the very different behaviour ofthe two mutants Tyr-159 -+ Ser and Tyr-159 -+ Phe. These resultsunderline the dangers of forming hypotheses on the basis of theproperties of one single mutant and interactions with only onesubstrate and one f-lactam.

This work was supported in part by the Belgian Government in the frame of the P6Ied'Atraction Interuniversitaries (PAI no. 19), an Action Concertee with the BelgianGovernment (convention 89/94-130), the Fonds de la Recherche ScientifiqueMedicale (contrat no. 3.4537.88), and a convention tripartite between the R6gionWallone, SmithKline Beecham U.K., and the University of Liege. We are indebted tothe Conseil de la Recherche (Universit6 de Liege) and the Fonds National de laRecherche Scientifique (F.N.R.S.) for grants that provided the stopped-flow and thequenched-flow apparatus (F.R.F.C. contrat no. 2.4503.90). J.-M.W. and M.J. areFellows of the Institut pour l'Encouragement de la Recherche Scientifique dansl'Industrie et l'Agriculture (I.R.S.I.A., Brussels, Belgium). B.J. and C.D. are ChercheursQualifies of the F.N.R.S.

REFERENCESAdam, M., Damblon, C., Plaitin, B., Chritiaens, L. and Frere, J.-M. (1990) Biochem. J.

270, 525-529Adam, M., Damblon, C., Jamin, M., Zorzi, W., Dusart, V., Galleni, M., El Kharroubi, A.,

Piras, G., Spratt, B. G., Keck, W., Coyette, J., Ghuysen, J.-M., Nguyen-Disteche, M. andFrere, J.-M. (1991) Biochem. J. 279, 601-604

Bourguignon-Bellefroid, C., Wilkin, J.-M., Joris, B., Aplin, R. T., Houssier, C., Prendergast,F. G., Van Beeumen, J., Ghuysen, J.-M. and Frere, J.-M. (1992) Biochem. J. 282,361-367

Christensen, H., Martin, M. T. and Waley, S. G. (1990) Biochem. J. 266, 853-861De Meester, F., Joris, B., Reckinger, G., Bellefroid-Bourguignon, C., Frere, J.-M. and Waley,

S. G. (1987) Biochem. Pharmacol. 36, 2393-2403Duez, C., Piron-Fraipont, C., Joris, B., Dusart, J., Urdea, M., Martial, J., Frere, J.-M. and

Ghuysen, J.-M. (1987) Eur. J. Biochem. 162, 509-518Erpicum, T., Granier, B., Delcour, M., Lenzini, M. V., Nguyen-Disteche, M., Dusart, J. and

Frere, J.-M. (1990) Biotechnol. Bioeng. 35, 719-726Frere, J.-M. and Joris, B. (1985) CRC Crit. Rev. Microbiol. 11, 299-396Frere, J.-M., Ghuysen, J.-M., Perkins, H. R. and Nieto, M. (1973) Biochem. J. 135,

483-492Frere, J.-M., Ghuysen, J.-M. and Perkins, H. R. (1975) Eur. J. Biochem. 57, 353-359Frere, J.-M., Duez, C., Ghuysen, J.-M. and Vandekerckhove, J. (1976a) FEBS Lett. 70,

257-260Frere, J.-M., Ghuysen, J.-M., Vanderhaeghe, H., Adriaens, P., Degelaen, J. and DeGraeve, J.

(1976b) Nature (London) 260, 451-454Frere, J.-M., Nguyen-Disteche, M., Coyehte, J. and Joris, B. (1992) in The Chemistry of

Beta-Lactams (Page, M. I., ed.), pp. 148-197, Chapman & Hall, LondonGhuysen, J.-M. (1991) Annu. Rev. Microbiol. 45, 37-67Hadonou, A. M., Jamin, M., Adam, M., Joris, B., Dusart, J., Ghuysen, J.-M. and Frere,

J.-M. (1992) Biochem. J. 282, 495-500Hopwood, D. A., Bibb, M. J., Chater, K. F., Kieser, T., Bruton, C. J., Kieser, H. M., Lydiate,

D. J., Smith, G. P., Ward, J. M. and Schremp, F. (1985) Genetic Manipulation ofStreptomyces: A Laboratory Manual, John Innes Foundation, Norwich

Jacob, F., Joris, B., Dideberg, O., Dusart, J., Ghuysen, J.-M. and Frere, J.-M. (1990a)Protein Eng. 4, 79-86

Jacob, F., Joris, B., Lepage, S., Dusart, J. and Frere, J.-M. (1990b) Biochem. J. 271,399-406

Jamin, M., Adam, M., Damblon, C., Chritiaens, L. and Frere, J.-M. (1991) Biochem. J. 280,499-506

Joris, B., Ghuysen, J.-M., Dive, G., Renard, A., Dideberg, 0., Charlier, P., Frere, J.-M.,Kelly, J., Boyington, J., Moews, P. and Knox, J. (1988) Biochem. J. 250, 313-324

Joris, B., Ledent, P., Dideberg, O., Fonze, E., Lamotte-Brasseur, J., Kelly, J. A., Ghuysen,J.-M. and Frere, J.-M. (1991) Antimicrob. Agents Chemother. 35, 2294-2301

Kelly, J. A., Dideberg, 0., Charlier, P., Wery, J., Libert, M., Moews, P., Knox, J., Duez, C.,Fraipont, C., Joris, B., Dusart, J., Frere, J.-M. and Ghuysen, J.-M. (1986) Science 231,1429-1 431


Kelly, J. A., Waley, S. G., Adam, M. and Frere, J.-M. (1992) Biochim. Biophys. Acta 1119,256-260

Lamotte-Brasseur, J., Dive, G., Dideberg, O., Charlier, P., Frere, J.-M. and Ghuysen, J.-M.(1991) Biochem. J. 279, 213-221

Leatherbarrow, R. J. (1987) ENZFITTER, Elsevier Biosoft, Cambridge, U.K.Malhotra, K. T. and Nicholas, R. A. (1992) J. Biol. Chem. 267, 11386-11391Maniatis, T., Fritsch, E. F. and Sambrook, J. (1982) Molecular Cloning: A Laboratory

Manual, Cold Spring Harbor Laboratory Press, Cold Spring HarborNieto, M., Perkins, H. R., Frere, J.-M. and Ghuysen, J.-M. (1973) Biochem. J. 135,

493-505

Oefner, C., Darcy, A., Daly, J. J., Gubernator, K, Charnas, R. L., Heinze, I., Hubschwerlen,C. and Winkler, F. K. (1990) Nature (London) 343, 284-288

Samraoui, B., Sutton, B., Todd, R., Artymiuk, P., Waley, S. G. and Phillips, D. (1986)Nature (London) 320, 378-380

Taylor, J. W., Ott, J. and Eckstein, F. (1985) Nucleic Acids Res. 13, 8765-8784Varetto, L. (1991) Ph.D. Thesis, Universit6 de Liege, BelgiumVaretto, L., Frere, J.-M., Nguyen-Disteche, M., Ghuysen, J.-M. and Houssier, C. (1987) Eur.

J. Biochem. 162, 525-531Waley, S. G. (1992) in The Chemistry of Beta-Lactams (Page, M. I., ed.), pp. 198-228,

Chapman & Hall, London

Received 15 October 1992/16 November 1992; accepted 20 November 1992

streptomyces r61 dd-peptidase

Documents